<p dir="ltr">During the past century automation has become an integral part of all facets of every aspect of life including robotics. There are several types of autonomous based systems; navigation, human interaction, decision making, controls, and others. Autonomously navigating robots are used in many applications including assistive robotics, military, space exploration, manufacturing, etc. Unmanned ground vehicles (UGV) are an example of autonomous systems falling under the category of navigation, where navigation is dominantly composed of automatic transport and movement in real world environments. Autonomous UGVs are highly used in hazard zones, search and rescue scenarios, and exploration endeavors that keep operators away from harm. Simultaneous Localization and Mapping (SLAM) provides the best approach to the problems faced in unknown environments. Visual based cameras, light detection and ranging (LiDAR) sensors, global positioning systems, and inertial measuring units (IMU) feed constant streams of data to the SLAM algorithm. These sensors allow the UGV to explore an unknown outdoor environment whilst traversing around obstacles. The objects focused in this research are construction barrels, cones, ramps, flags, and white lanes, while the unknown environment typically tested is open field. Utilizing open source robot operating system ROS packages, the UGV navigation algorithm uses 2D LiDAR, IMU, GPS, and depth camera data to combine the sensor inputs for a reliable robotic system. This paper demonstrates a robotic system combining RGB depth camera, object detection, and conventional outdoor SLAM navigation in unknown environments.&nbsp;</p><p dir="ltr">The developed UGV control algorithm uses ROS as a middleware between sensor input, SLAM, object detection, and navigation. The 2D occupancy grid map is produced by OpenSlam’s gmapping using an RP LiDAR, but is edited before navigation, to incorporate key visual data that would otherwise be missed without a camera. Environments that may lack a consistent stream of obstacles for pure LiDAR localization benefit from less reliable, but drift independent, GPS localization. GPS assisted SLAM and waypoint creation allows for reduced error propagation when lacking reliable odometry or 3D obstacles for lidar localization. Using the waypoint_nav package, GPS coordinate goals are inputted for the UGV to navigate without the need for teleoperation. These waypoints are presented one at a time or as a list of coordinates for subsequent navigation once a waypoint is reached. The stereoscopic depth camera's main purpose is taking the video output and detecting lanes a LiDAR would otherwise miss. The lane detection algorithm puts the raw output through a series of filters which removes shadows, combines color channels, and collects point data. Once detected they are then published on to the cost map which is used by the planner for path generation. Ramps or slopes that the LiDAR would deem impassable are overruled through the use of object detection and the specialized navigation algorithms are employed around a specific object. Once a ramp or flag has been detected an alternative path is implemented that is different from the global GPS path. The UGV combines all these methods to reliably explore unknown environments without the need for teleoperation.</p>

Navigation and Path Planning of an Autonomous Mobile Robot

Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2021/volume/85543

ASME 2021 International Mechanical Engineering Congress and Exposition (IMECE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.